Einstein–de Sitter Universe Model

• The Einstein–de Sitter universe is a flat, matter-dominated cosmological model defined by zero cosmological constant and spatial curvature.
• Its background dynamics yield a scale factor proportional to t^(2/3), establishing a direct relation between cosmic expansion and critical density.
• The model serves as a benchmark in cosmology, offering an analytically tractable framework for testing perturbation theory and general relativity on large scales.

The Einstein–de Sitter universe is the flat, matter-dominated relativistic cosmological model with zero cosmological constant and zero spatial curvature: $k=0$, $\Lambda=0$, $p=0$, and $\rho=\rho_c$ at all times. In modern notation it is the unique dust-dominated FRW model with $a(t)\propto t^{2/3}$, and historically it became the benchmark “big bang” model for much of the twentieth century because it supplied a simple relation between cosmic expansion and mean density, $H^2=(8\pi G/3)\rho$ (O'Raifeartaigh et al., 2020, O'Raifeartaigh et al., 2015).

1. Definition and spacetime structure

In modern notation, the Einstein–de Sitter model is obtained by specializing the Robertson–Walker metric to the spatially flat case,

$ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$

or, in comoving spherical coordinates,

$$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$

Its matter content is a pressureless fluid (“dust”), with energy–momentum tensor $T_{\mu\nu}=\rho\,u_\mu u_\nu$, and its defining assumptions are $k=0$, $\Lambda=0$0, and $\Lambda=0$1 (O'Raifeartaigh et al., 2015).

Within the flat FLRW ansatz, Einstein’s field equations,

$\Lambda=0$2

reduce to the Friedmann system

$\Lambda=0$3

For the Einstein–de Sitter case, all other components such as radiation and $\Lambda=0$4 are set to zero [(O'Raifeartaigh et al., 2015); (Baker, 2011)].

A normalized form used in recent mathematical work writes the metric on $\Lambda=0$5 as

$\Lambda=0$6

with

$\Lambda=0$7

This formulation makes explicit that the Einstein–de Sitter spacetime is spatially homogeneous and isotropic and undergoes decelerated expansion (Bernhardt et al., 10 Jul 2026).

2. Exact background dynamics

The continuity equation for pressureless matter,

$\Lambda=0$8

implies $\Lambda=0$9. Substituting this relation into the Friedmann equation yields the standard power-law solution

$p=0$0

Equivalently, Einstein’s 1933 review gives

$p=0$1

with $p=0$2 fixed by $p=0$3 or by normalizing $p=0$4 at the present time [(O'Raifeartaigh et al., 2015); (Baker, 2011)].

The matter density then obeys

$p=0$5

so that cosmic time and density are directly related by

$p=0$6

For the present epoch, the model gives

$p=0$7

Using $p=0$8, one also obtains

$p=0$9

These relations are among the main reasons the Einstein–de Sitter model remained analytically important: background evolution is closed-form throughout [(O'Raifeartaigh et al., 2020); (Baker, 2011)].

The critical density is

$\rho=\rho_c$0

and in the Einstein–de Sitter case one has $\rho=\rho_c$1 exactly. In the more general FRW setting, departures from $\rho=\rho_c$2 imply $\rho=\rho_c$3 (O'Raifeartaigh et al., 2015).

3. Einstein and de Sitter’s 1932 construction

In their 1932 paper, Einstein and de Sitter considered a relativistic model of the expanding universe with both the cosmological constant and the curvature of space set to zero. Their original notation wrote the flat line element as

$\rho=\rho_c$4

and, for $\rho=\rho_c$5 and $\rho=\rho_c$6, they quoted the equation

$\rho=\rho_c$7

with $\rho=\rho_c$8 and $\rho=\rho_c$9. Multiplying by $a(t)\propto t^{2/3}$0 recovers the modern form

$a(t)\propto t^{2/3}$1

(O'Raifeartaigh et al., 2020).

They also introduced two characteristic lengths, $a(t)\propto t^{2/3}$2 and $a(t)\propto t^{2/3}$3, through

$a(t)\propto t^{2/3}$4

so that

$a(t)\propto t^{2/3}$5

Inserting Hubble’s then-accepted value $a(t)\propto t^{2/3}$6 led to numerical estimates of $a(t)\propto t^{2/3}$7, $a(t)\propto t^{2/3}$8, and $a(t)\propto t^{2/3}$9 (O'Raifeartaigh et al., 2020).

Historically, the paper emerged during a short Caltech collaboration in 1932. Einstein had already abandoned $H^2=(8\pi G/3)\rho$0 in his 1931 closed-universe model, and Heckmann had pointed out that $H^2=(8\pi G/3)\rho$1 need not be positive. Einstein and de Sitter then selected the $H^2=(8\pi G/3)\rho$2, $H^2=(8\pi G/3)\rho$3 case because it yielded a testable relation between expansion rate and $H^2=(8\pi G/3)\rho$4 without the extra “free parameter” of curvature. Although Friedman and Lemaître had already shown that non-static $H^2=(8\pi G/3)\rho$5 models exist, the 1932 note delivered the first explicit dynamic solution with $H^2=(8\pi G/3)\rho$6, $H^2=(8\pi G/3)\rho$7 (O'Raifeartaigh et al., 2020).

The paper’s brevity became part of its historical reputation. The mathematics are described as puzzling to modern eyes, and the authors do not explicitly consider the evolution of the cosmos in the note itself. At the same time, the model’s simplicity and specificity made it an important benchmark for both theorists and observers (O'Raifeartaigh et al., 2020).

4. Critical density, $H^2=(8\pi G/3)\rho$8, and its role as the standard big-bang prototype

The density parameter is defined by

$H^2=(8\pi G/3)\rho$9

For the Einstein–de Sitter case, $ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$0 with matter only. In the standard classification summarized in the historical literature, $ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$1 corresponds to a closed universe that eventually recollapses, $ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$2 to an open universe expanding forever, and $ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$3 to the balanced Einstein–de Sitter case (O'Raifeartaigh et al., 2020).

This framework made the model central to observational cosmology. The Einstein–de Sitter density became the “critical density” against which astronomers could compare observations, and it provided theoreticians with a dividing line between closed and open models. Observational programmes of the 1950s–70s, often characterized as the “search for two numbers,” $ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$4 and $ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$5, used $ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$6 as the default hypothesis for a big-bang cosmos (O'Raifeartaigh et al., 2020).

Einstein’s 1933 review reinforces the same logic of parsimony. By late 1932 there was no observational evidence for either a nonzero $ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$7 or spatial curvature, and Einstein’s stated aim was the simplest model of the cosmos that could account for observation. In his formulation, once expansion is admitted, “it is no longer necessary to introduce the universal constant $ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$8,” and “the fact of a non-zero density of matter need not be reconciled with a curvature of space, but instead with an expansion of space” (O'Raifeartaigh et al., 2015).

The model nevertheless carried numerical tensions when evaluated with early data. Using $ds^2=-c^2dt^2+a^2(t)[dx^2+dy^2+dz^2],$9, one finds $$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$0 years in exact arithmetic, whereas Einstein quoted “$$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$1 years,” probably as a rough upper bound. This episode is historically significant because it shows that the Einstein–de Sitter framework was simple and predictive, but also vulnerable to the quality of the available measurements (O'Raifeartaigh et al., 2015).

5. Linear perturbations and use as a testbed

Because the Einstein–de Sitter universe is spatially flat, contains only pressureless matter, and has no cosmological constant, it remains an analytically tractable “toy model” for perturbation theory and for model-independent tests of General Relativity on cosmological scales (Baker, 2011).

In the Newtonian (longitudinal) gauge, neglecting pressure and anisotropic stress, the density contrast $$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$2 satisfies

$$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$3

Specializing to the Einstein–de Sitter background, where $$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$4 and $$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$5, one obtains two linearly independent solutions: $$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$6 Thus the growing mode scales exactly as the scale factor in General Relativity (Baker, 2011).

On superhorizon scales, the curvature perturbation on uniform-density hypersurfaces,

$$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$7

is conserved provided there are no entropy perturbations. In pure Einstein–de Sitter, both metric potentials $$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$8 and $$ds^2=-c^2dt^2+a^2(t)\,[dr^2+r^2(d\theta^2+\sin^2\theta\,d\phi^2)].$$9 remain constant outside the horizon, so $T_{\mu\nu}=\rho\,u_\mu u_\nu$0 is constant. Since the Integrated Sachs–Wolfe contribution is proportional to

$T_{\mu\nu}=\rho\,u_\mu u_\nu$1

constant $T_{\mu\nu}=\rho\,u_\mu u_\nu$2 and $T_{\mu\nu}=\rho\,u_\mu u_\nu$3 imply no ISW contribution during a pure matter era (Baker, 2011).

This same background is widely used as a safe testbed for parameterized gravity. In one common parameterization,

$T_{\mu\nu}=\rho\,u_\mu u_\nu$4

while a more field-equation-consistent form allows

$T_{\mu\nu}=\rho\,u_\mu u_\nu$5

On subhorizon scales the density-contrast equation then generalizes so that the power-law ansatz $T_{\mu\nu}=\rho\,u_\mu u_\nu$6 gives modified exponents $T_{\mu\nu}=\rho\,u_\mu u_\nu$7. In the GR limit $T_{\mu\nu}=\rho\,u_\mu u_\nu$8, $T_{\mu\nu}=\rho\,u_\mu u_\nu$9, the growing-mode index returns to $k=0$0 (Baker, 2011).

6. Philosophical significance, limitations, and early-universe caution

The Einstein–de Sitter model was not only a technical construction but also an expression of a philosophical preference for simplicity. By excising both the cosmological constant and spatial curvature, Einstein and de Sitter embodied a preference for the simplest theory consistent with observation. The model also challenged the older view, associated with Mach’s ideas, that relativistic inertia required a closed universe with $k=0$1: in Einstein–de Sitter, an infinite Euclidean cosmos can carry a finite mean density and still expand (O'Raifeartaigh et al., 2020).

Einstein’s 1933 review makes the same point in a more explicit form. By 1932 he had abandoned the static universe, then $k=0$2, and then $k=0$3, because there was no empirical need for those ingredients. This gave the model its logical economy, but not unrestricted applicability (O'Raifeartaigh et al., 2015).

A recurrent misconception is that the Einstein–de Sitter model was intended as a complete description of all cosmic epochs. The historical record does not support that reading. Einstein recognized that extrapolating homogeneous dust back to the origin leads to a beginning in which $k=0$4, and he explicitly noted that the approximation of spatially uniform density breaks down at such early times. The stated reason was that the rough approximation “according to which the density $k=0$5 is independent of location” cannot be trusted there; spatial inhomogeneities, radiation pressure, and other physics must enter (O'Raifeartaigh et al., 2015).

A second misconception concerns originality. Although the 1932 paper was neither the first relativistic expanding-universe paper nor a full treatment of non-static cosmology, the historical analysis concludes that it was unique in providing the first specific analysis of the dynamic $k=0$6, $k=0$7 case and in establishing a straightforward relation between expansion and mean density. This suggests that its importance lay less in formal novelty than in fixing a benchmark model with immediate observational use (O'Raifeartaigh et al., 2020).

7. Stability and later reinterpretations

The Einstein–de Sitter universe continues to function as a reference solution in mathematical cosmology. Recent work describes it as the prevailing model used in cosmology to describe the cold dark matter-dominated epoch of the universe, while also emphasizing a subtle instability structure: under the Einstein–Euler equations with a pressureless fluid equation of state, the model is linearly unstable, but with a polytropic equation of state

$k=0$8

the nonlinear dynamics are different (Bernhardt et al., 10 Jul 2026).

For $k=0$9 and Sobolev regularity $\Lambda=0$00, there exists an open set of initial data on $\Lambda=0$01 with metric near $\Lambda=0$02, second fundamental form near $\Lambda=0$03, fluid density near a positive constant, and velocity near zero, such that the gauge-fixed Einstein–Euler evolution is future-global and geodesically complete, and converges to an Einstein–de Sitter spacetime up to a linear change of spatial basis. The proof uses conformal rescaling to expansion-normalized variables so that the background Einstein–de Sitter solution becomes exactly Minkowski, together with a sourced wave gauge and corrected energy estimates. Physically, the result shows that adding an arbitrarily small positive pressure compatible with cold dark matter yields a fully nonlinear global attractor (Bernhardt et al., 10 Jul 2026).

A distinct contemporary line of work proposes an inhomogeneous Einstein–de Sitter (iEdS) universe. In that framework, the usual strictly homogeneous FRW ansatz is replaced by an ensemble of disjoint regions with local scale factors $\Lambda=0$04, and the global scale factor is defined by volume averaging. The resulting global deceleration parameter contains a variance term, so global acceleration can occur even if each cell decelerates. The proposal assumes pressureless matter and curvature only, with no dark energy, and argues that effective negative curvature emerges dynamically from growing inhomogeneities without breaking spatial flatness (Raffai et al., 5 Nov 2025).

In this iEdS picture, the global evolution transitions quasilinearly from Einstein–de Sitter, $\Lambda=0$05, to a Milne state, $\Lambda=0$06. Two realizations, iEdS(1) and iEdS(2), are reported to fit CMB, BAO, and SN Ia data while alleviating or resolving the $\Lambda=0$07 tension, with both models yielding $\Lambda=0$08. These constructions are not the canonical Einstein–de Sitter universe; rather, they show that the Einstein–de Sitter background remains a live organizing principle in attempts to reinterpret late-time acceleration through inhomogeneity and averaging rather than $\Lambda=0$09 (Raffai et al., 5 Nov 2025).